The Construction-Integration model has suggested that written discourse comprehension involves the construction of a propositional text base from which an integrated and coherent mental representation or situation model is formed. This requires readers to go beyond written words and sentences and make inferences on the discourse level to form a coherent understanding of the text. Previous research has suggested that individual differences in vocabulary and working memory explain unique variance in inferencing in the English language. However, less well understood is whether these factors play a similar role in inference making in a non-alphabetic orthography, such as Chinese characters. Given the phonological, morphological and orthographic differences between the English and Chinese languages/orthographies, better understanding of the processes that contribute to Chinese inference making is of theoretical and practical importance for Chinese readers.

Studies 1 and 2 examined the contributions of vocabulary and working memory to the generation of inferences when reading Chinese texts. In Study 1, Chinese oral vocabulary, working memory (backward digit span, reading span, nonword repetition and backward spatial span) and Chinese inference making measures were administered to 65 Mandarin– speaking children in Grades 5 and 6. The Chinese inference making measure involved children reading narrative and expository passages and answering text-connecting and knowledge-based inferences on the texts. Correlational analyses found that Chinese vocabulary and verbal working memory (Chinese backward digit span, reading span and nonword repetition) were associated with Chinese inference making. Multiple hierarchical regression analyses revealed that Chinese vocabulary and verbal working memory, as measured by the backward digit span task, were found to uniquely predict inference-making scores produced by this cohort. Study 2 further explored the contributions of Chinese vocabulary and verbal working memory to inference making after controlling for the Chinese character word reading ability. Chinese oral vocabulary, character word reading, verbal working memory and inference making measures were administered to 60 children in Grades 5 and 6. Multiple hierarchical regression analyses revealed similar results to Study 1, in that Chinese vocabulary and backward digit span tasks made unique contributions to inference making in this cohort. However, after controlling for character word reading, Chinese vocabulary was not a significant predictor of inference making; Chinese character word reading and backward digit span tasks were found to explain significant variance in inference making. Separate analyses for each grade suggested different patterns: after controlling for character word reading, Chinese vocabulary and backward digit span tasks uniquely predicted inference making in Grade 5, whereas none of the factors was found to predict inference making in Grade 6.

Study 3 investigated the contributions of verbal working memory and vocabulary to inference making among a group of Chinese–English bilingual children in Grade 8. The study examined potential influences of first language Chinese predictors on second language English inference making. Fifty-five eighth graders who used Mandarin as their first language and English as a second language completed parallel tests of verbal working memory, vocabulary and inference making in both Chinese and English. Participants performed significantly better in the Chinese measures. Regression analyses revealed that Chinese vocabulary uniquely predicted Chinese inference making but neither English vocabulary nor English verbal working memory predicted English inference making. Cross- language analyses suggested that Chinese vocabulary was a significant predictor of English inference making.

The aim of the three studies was to inform theories of inference making in Chinese children and the findings have practical implications to educators in highlighting the importance of nurturing children’s Chinese vocabulary and word reading abilities to facilitate inference making during reading. The findings also suggest that further studies of the phonological and central executive systems of working memory would be useful in determining those factors influencing inferential comprehension, particularly for the Chinese fifth and sixth graders. Furthermore, the finding that Chinese vocabulary knowledge uniquely predicted bilingual eighth graders’ English inferential processing argues for the need to consider first language competency when investigating bilingual reading abilities.